Locally orderless networks

arXiv:2406.13514v1 Announce Type: new
Abstract: We present Locally Orderless Networks (LON) and its theoretic foundation which links it to Convolutional Neural Networks (CNN), to Scale-space histograms, and measurement theory. The key elements are a regular sampling of the bias and the derivative of the activation function. We compare LON, CNN, and Scale-space histograms on prototypical single-layer networks. We show how LON and CNN can emulate each other, how LON expands the set of functionals computable to non-linear functions such as …

abstract arxiv bias cnn convolutional convolutional neural networks cs.cv elements foundation function histograms key layer links measurement networks neural networks sampling scale space the key theory type